Feminist solidarity building as embodied agonism: An ethnographic account of a protest movement

Feminist solidarity, after early and idealistic conceptions of an all-encompassing sisterhood, has become preoccupied with understanding and theorizing differences between women. This study develops an account of solidarity as embodied agonism, where difference and contest are experienced and negotiated through the body. Difference and contest are reframed within feminist solidarity projects as resources for, rather than inhibitors to, generating collective agency. This is done through an ethnography of a protest movement in Montenegro, which drew together diverse groups of women, and bring our data into conversation with theories of agonistic democratic practice and embodied performativity. Embodied agonistic solidarity is theorized as a participative and inclusive endeavour driven by conflictual encounters, constituted through the bodies, language and visual imagery of assembling and articulating subjects. Our account of solidarity is presented as constituted through three dimensions, each of which represents a different emphasis on sensory experience: exposing , which is to make one's body open to the hardship of others, enabling alliances between unlikely allies to emerge; citing , which is to draw on others ’ symbolic resources and to publicly affirm them; inhabiting , which is to embody the deprivations of others, enabling alliances to grow and persist.


| INTRODUCTION
Solidarity building has long preoccupied feminist academics and practitioners invested in the notion that significant social change can only be achieved collectively (Arruzza, Bhattacharya, & Fraser, 2019).Yet with some notable exceptions (e.g., Cranford, 2012;Vacchani & Pullen, 2019), feminist solidarity building remains under-theorized within organization studies, meaning that we do not know enough about the practices that can assist its generation.
Conceptually, feminist solidarity and sisterhood have lost some potency in recent decades, as the development of poststructuralist theory and lessons gleaned from practical experience have legitimately questioned whether 'women' can be 'essentialized' as a homogeneous category.An emphasis on difference over unity started to become a lived reality when women worldwide realized that 'our herstories are too numerous and too varied' to be taken as 'shared' (Baker & Diawara, 1996, p. 62).
This study draws on ethnographic data from a major women's protest and proposes a democratic form of embodied agonistic solidarity.Our main contribution lies in offering a framework to analyse how women work together, drawing on difference as a resource offering 'agonistic' energy and vitality (Mouffe, 2009a(Mouffe, , 2013)).Embodied agonistic solidarity is a practice that cultivates collective agency through difference and conflict.We develop our account by placing Chantal Mouffe's (2009aMouffe's ( , 2009bMouffe's ( , 2013Mouffe's ( , 2014) ) theory of agonistic democratic practice and Judith Butler's (2006Butler's ( , 2011Butler's ( , 2015Butler's ( , 2016) ) theory of performativity in dialogue with data generated through an ethnographic study in Montenegro.
The study offers insight into solidarity building practice within a diverse anti-government protest movement in Montenegro centred on overturning a decision to curtail and revoke benefits for mothers.We argue that solidarity is a contested practice, where difference is experienced and negotiated through corporeal-discursive enactment.
Drawing on agonistic notions of democratic practice we show how the collective agency of women develops as embodied solidarity, involving both positive expressions of empathy, vulnerability and care, but also contest, pain and discomfort.From this basis we propose three dimensions of embodied agonistic solidarity.First, exposing means making one's body open to the hardship of others, enabling alliances between unlikely allies to emerge.Second, citing denotes the drawing in of others' symbolic resources and publicly affirming them.Third, inhabiting involves living through, in embodied and intimate ways, the deprivations of others, which strengthens solidarity and enables alliances to grow and persist.
The article is structured as follows.First, we review the literature on feminist solidarity building, emphasizing the preoccupation with theorizing difference.We then propose an embodied performative and democratic conceptual framework as a means of exploring the enactment of solidarity in practice, before introducing our ethnographic and multimodal methodology, as well as the description of our research setting.Finally, we present our three dimensions of embodied agonistic solidarity, before discussing their relevance and contribution in relation to the literature.

| FEMINIST SOLIDARITY BUILDING
Solidarity has been recognized for its power in fighting oppression, strengthening resistance, instigating social change and binding movements together (Allen, 1999;Federici, 2009;hooks, 1984;Mohanty, 2003).Initially, it was envisaged as a value that emerged from shared experiences of sexist oppression (Bell, Meriläinen, Taylor, & Tienari, 2019;Božinovi c, 1996).'Woman' was a universal subject and by implication it was believed that all women could be fused in solidarity against patriarchy regardless of political affiliation, sexual orientation, class, race and other social categories (Hemmings, 2012;Sen, 2019;Sweetman, 2013).
A counter-narrative emerged, as scholars and practitioners began questioning the ability of a universal form of solidarity to capture all forms of oppression women face globally under a single category (Allen, 1999;Yuval-Davis, 2011).For hooks (1984) the notion of an all-encompassing sisterhood was a bourgeois project that failed to address the complex and variegated nature of women's experiences and forms of oppression.Moreover, the 'whitewashed ' (Bell et al., 2019, p. 7) conception of feminist solidarity served to conceal the multifaceted and diverse experiences of being a woman across different times and geographical spaces (Mohanty, 2003), not to mention the 'exclusion of some women in social justice movements' (Vacchani & Pullen, 2019, p. 27).Some of these main 'fault-lines' emerged through race, ethnicity, gender, class, nationality and sexuality (Sen & Ratan, 2006, p. 5; see also Blasius, 1994;Connell, 2012;Franzway & Fonow, 2008;hooks & McKinnon, 1996;Thornton-Dill, 1983).One solution to the recognition of important differences and exclusions was to examine the particularities of experience with greater care, thereby countering the ideological postulate of the universal 'woman' (Mohanty, 1991).These scholars insisted that the theorizing of 'women' ought to pay attention to the particular contexts women were situated within, embracing a heterogeneous, historically, socially and politically nuanced perspective (Chant, 2010;Sen & Grown, 1987).
This preoccupation with difference drove the emergence and continuation of diverse conceptual forms of feminisms (Bell et al., 2019;Özkazanç-Pan, 2018).One of the most prominent, intersectionality, aims to avoid the reification of experiences (Bonfiglioli, 2016;Cole, 2008;Crenshaw, 1991;Yuval-Davis, 2011) by highlighting oppression occurring at the intersection of various categories of social identity (e.g., woman, black, lesbian, etc.).Such understandings can inform a sensitivity to important issues of difference and asymmetrical power in the forging of solidarity.
Neoliberalism has provided further challenges to solidarity projects by intensifying the tendency towards individualism and away from collective modes of thought and action (Arruzza et al., 2019;Vacchani & Pullen, 2019).It has, and continues to, appropriate forms of feminism for its own commodified agenda, where women's aspirations, interests and capacities are mobilized for market ends (Hopkinson & Aman, 2017) and 'women's empowerment [is conflated with] neoliberal reason' (Chatterjee, 2019, p. 378).
Yet resistance to neoliberalism has emerged as a potential unifying cause as increasing numbers of people experience precarity in and out of work (Arruzza et al., 2019;Federici, 2009).Neoliberalism, in favouring private profit at the expense of social infrastructure and the environment (Ayers, 2006;Sahoo, 2014), creates common experiences of oppression (Mohanty, 2003), and affective, embodied solidarity has been recognized as a response to these adverse effects (Vacchani & Pullen, 2019).
Nevertheless, extant literature tends to downplay the potential of difference as a resource for building feminist solidarity, instead focusing on either particular or common experiences.The remainder of the article offers a conceptual framework that reads solidarity in a manner that seeks to reframe differences within feminist solidarity movements as a strength.

| EMBODIED AGONISTIC SOLIDARITY: A CONCEPTUAL FRAMEWORK
In this section we place Mouffe's (2009aMouffe's ( , 2009bMouffe's ( , 2013Mouffe's ( , 2014) ) concept of agonistic democratic practice in dialogue with Butler's (2006Butler's ( , 2011Butler's ( , 2015Butler's ( , 2016) ) account of embodied performativity to interpret the enactment of feminist solidarity in practice.Interweaving these can help overturn some of the common misconceptions regarding solidarity building (i.e., that difference and conflict inhibit solidarity).Further, it can potentially connect and recognize the inseparability of speech and body acts, foregrounding the materiality of language and the discursivity of corporeality, and thereby surface the nuanced enactment of difference.This is an important concern for a feminism that seeks to illuminate the complex and intricate nature of women's lives, enmeshed in patriarchal norms but where such norms are often experienced and resisted through the body (Butler, 2016).
Agonists maintain that conflict is ineradicable in any pluralist society, for there will always be an 'us' and 'them' amongst whom differences cannot be finally resolved (Mouffe, 2013;Rhodes & Wray-Bliss, 2012).Since plurality is enacted as different and/or opposing democratic identities that might stand in antagonistic relation to one another (Mouffe, 2009a), rather than trying to tame this, we should embrace and deploy it to invigorate democratic practice (Mouffe, 2009a(Mouffe, , 2013;;Rhodes & Harvey, 2012).

SMOLOVI
C JONES ET AL.
Central to this foregrounding of contestation is the idea of a conflict of 'adversaries', rather than 'enemies' (Mouffe, 2009a, p. 13).While each 'adversarial' side will seek to establish its own 'hegemonic order', that does not necessarily mean that they are 'enemies', as they share a common allegiance to the democratic principles of 'liberty and equality for all' (Mouffe, 2013, p. 7), although they may disagree about the meaning of these principles in practice.In agonism, each group will aim to establish its own understanding of democratic principles via a 'vibrant democracy' (Mouffe, 2013, p. 7) of continuous debate.
What is radical about this view is the recognition that although some agreement can be reached on democratic principles, a core of negativity will always inhere in any agonistic relation, as no system of language is ever capable of capturing all meaning satisfactorily (Mouffe, 2018); while proponents hope that agonism may act to channel people away from antagonism, they hold that no form of democratic practice can, or should seek to, overcome the foundational negativity within language that prevents full closure of meaning (Mouffe, 2018).Conflict will always be present, and should be encouraged, as no worldview can provide unity to an identity or society (Smolovi c Jones, Smolovi c Jones, Winchester, & Grint, 2016).
Indeed, identity is at the heart of democratic practice for agonists and is something conceptualized as relational and dependent on an antagonistic outside: the hegemonic constellation of 'us' stands in adversarial relation to a hegemonic constellation of 'them' (Mouffe, 2013(Mouffe, , 2018) ) the relation here is not defined in terms of reified commonality but as a structural condition of democratic practice towards a dynamic substantive 'them' always in the process of formation and re-formation.This point reinforces the conflictual heart of Mouffe's theory in that any democratic practice will always entail exclusions, as various groups fight to assert a form of hegemony.Such an insight allows us to view the practice of building solidarity as something fluid and pluralistic, with the capacity for new alliances across difference forming, adapting and re-forming over time.It is a perspective that accepts that no group formed in solidarity willor shouldachieve consensus, but instead can harness conflict and difference in order to strengthen internal ties and develop as a political stance in relation to external aggressors (Vacchani & Pullen, 2019).
Mouffe's theory, however, concerns contestation primarily through language, which appears insufficient for understanding and accounting for an everyday form of feminist democratic practice that will stray beyond, but also include, language.Therefore, we favour a more embodied, affective reading (Butler, 2011(Butler, , 2015;;Hemmings, 2012;Tyler, 2019;Vacchani & Pullen, 2019).We turn primarily to Judith Butler's work on embodiment, assembly and vulnerability to help us understand how democratic forms of solidarity building may be experienced and constituted through the body as well as through language.
An embodied focus offers an understanding of how subjects can resist norms in unexpected ways, drawing on potent shared affects such as 'pleasure' and 'rage' (Hemmings, 2012, p. 150;Vacchani & Pullen, 2019) to craft more nuanced, relational practice/subjectivities. Butler (2011) affirms the existence of the material body that undermines and informs linguistic practices and is not completely 'intelligible' via language.This infers that language fails to entirely signify what it refers to via the term 'body' (Cabantous, Gond, Harding, & Learmonth, 2016;Fotaki, Metcalfe, & Harding, 2014); there will always be an excess of a signified referent that is 'never fully of language' (Butler, 2011, p. 37), a sense that bodies are ineffable, escaping our dominant systems of 'knowing'.
We enact solidarity practices through our bodies, by conveying, receiving and co-creating meaning (e.g., occupying streets in protests or blocking police cordons).Logocentrism ignores how speaking is a bodily act that requires 'larynx, the lungs, the lips and the mouth' (Butler, 2004, p. 172).Sound, digital, written and sign language are also material, relying on the body for communicative effect.We 'park' our bodies amongst others, in the street, workplace and social spaces, and in doing so bodies enact a certain meaning (Butler, 2015).Since the body conveys and enacts meaning through its presence, even when seemingly passive (Coupland, 2015) it is always performative.
Butler's account enables a view of solidarity building as dynamic and continuous entanglements of social/corporeal acts that reproduce and reconfigure one another through practice (Thanem & Knights, 2012).She conceptualizes such sites of democratic and solidarity practices as assemblies (Butler, 2015), which are embodied, collective and relational acts where subjects come together to assert and demand.Here, two of its key aspects are highlighted.
First is the notion of vulnerability (Butler, 2006(Butler, , 2015;;Tyler, 2019), which is an ontological condition of the humanthe basis for connection, understanding and solidarity.Bodies are necessarily vulnerable, that is, temporary vehicles that experience pain and death, which therefore form a basis for empathic understanding and solidarity.Vulnerability should not be mistaken for weaknessby contrast, it is constitutive of agency, as subjects learn to experience and build from this shared condition: 'receptivity and responsiveness become the basis for mobilising vulnerability rather than engaging in its destructive denial' (Butler, 2016, p. 25).People express a collective 'dissonance' between embodied experience and normative rhetoric (Vacchani & Pullen, 2019).
The second aspect is relational infrastructure.Moments where vulnerability is surfaced, for Butler (2015), underline the inherently relational condition of all subjectivity: "each 'I' brings the 'we' along" (p.51).Experiences of loss illuminate how any subject is dependent upon an 'infrastructure' (Butler, 2016, p. 21) of otherstheir physical presence, words and affective support.Once infrastructure is removed, we see in sharp focus our intra-dependencies on others and cultivating such awareness and experiences of relational dependence can therefore mobilize empathy for others' vulnerability (Vacchani & Pullen, 2019, p. 36).Understood in this way 'the body is less an entity than a relation, and it cannot be fully dissociated from the infrastructural and environmental conditions of its living' (Butler, 2016, p. 19).Generative experiences of assembly are capable of building a powerful form of collective agency because they create instances of people forming infrastructures of dependence between one another.
In sum, agonism offers a democratic frame for interpreting the process of feminist solidarity as one that can foreground conflict in a way that enacts and seeks to enshrine equal voice and participation, thus viewing difference as generative rather than inhibiting.Yet its focus on language underplays the role of the body in drawing people together in shared experience and assembly.Understanding agonistic engagement as shot through and driven by embodied relationality, where bodies as well as language 'speak', challenge and clash, offers a necessary additional frame for interpreting feminist solidarity.

| The research setting
The research setting was selected to enable a rich account of diverse people coming together in solidarity, despite seemingly incompatible differences.This ethnographic research took place over 11 months, during which the primary researcher assumed a role of participant-observer in the Women's Rights Centre (WRC) non-governmental organization (NGO) in Montenegro.WRC (2017) strives to establish 'gender-equal democratic practice' through advocacy, artful activism and the provision of free legal and psychological assistance.WRC is small (five permanent members of staff) but impactful, frequently quoted in the national media, and sought by numerous transnational organizations such as the United Nations and European Union for its views.Its focus covers policy advocacy and hands-on support for women suffering violence.
The research took place during a period of unrest in the country caused by the 'protest of mothers'.In 2016, the corruption-riven government (Williams, Radevi c, Gherhes, & Vorley, 2017) passed a law guaranteeing mothers of at least three children lifelong financial benefits if they renounced their jobs and pensions.Approximately 22,000 women opted for this incentive but were later left vulnerable when the governing party revoked the law. 1 The women affected were largely poor, tended towards more conservative and nationalist politics and were historically passive in terms of political engagement.After revocation however, thousands of mothers protested, occupying the squares of parliament and government and eventually participating in a hunger strike.
Between December 2016 and the time of writing there have been 36 street occupations and an ongoing online campaign.
In order to understand the significance of the protests it is important to underline the prevalence of a 'sanctity of motherhood' discourse in Montenegro, where the domestic and rearing identities of women are politically incentivized and socially cherished.Such a normative framework of motherhood shapes the experiences of women in work and family life but also expectations of what governments should and should not do in relation to the provision of services and benefits.WRC did not initially support the law on benefits as it was incongruent with its political stance; first, because such legislation incentivized women to remain in their homes instead of participating actively in the public sphere, and second, civil sector practitioners (henceforth referred to simply as 'practitioners') were cautious about allying themselves with protestors who explicitly drew on conservative patriarchal discourses of family and motherhood.However, having witnessed how many women depended on these benefits for basic sustenance and the way the government used mothers to obtain votes, WRC began exploring how it could support them, albeit unsure of the ways in which it could do so.The analysis offered in this article revolves around the process of solidarity building between these unlikely allies.

| Data generation
Ethnographic data was generated through first-hand engagement, where the first author gained insight into the protest movement through daily engagement with both practitioners and protesting mothers.Such an approach surfaced a rich account of the enactment of solidarity within a context of public corruption and patriarchy.The first author is a native Montenegrin familiar with the socio-political context through previous working experience in the country's civil sector, which helped her embed herself within WRC.Yet familiarity with the challenges addressed by WRC and her close proximity to it necessarily involved a partial perspective, which necessitated the need for deep self-reflection on the taken-for-granted views she might carry with her in interpreting the justness of various perspectives articulated in the field.Such self-reflection was not necessarily a solitary pursuit but a relational one, of inserting herself in places and with people who might challenge her, for example, more conservative mothers and other NGO workers with contrary views.
We pursued a discursive and multimodal form of ethnography (Cunliffe, 2010;Machin, 2016;Machin & Mayr, 2012), where knowledge is co-produced between researcher and participants through linguistic, material (e.g., images, posters, banners, placards, etc.) and embodied practices.Within data generation, following Butler (2011), the material and the linguistic were approached as entangled and inseparable.This meant refusing the binary between corporeal and discursive (Butler, 2011), instead viewing corporeal materiality and enunciated language as intra-dependent and indistinguishable.For example, wrinkles on a face might 'speak' of its history of hard labour.
We were interested in not only presenting an account of the generative practices of activists but also how they enacted such practices within a field of restrictive patriarchal norms (Ford, Harding, Gilmore, & Richardson, 2017).It was therefore important to explore the 'power of discourse to produce the phenomena that it regulates and constrains' (Butler, 2011, p. xii) by investigating the traces of a dominant patriarchal discourse in the language, artefacts and actions of activists; but we also explored 'the persistence of disidentification' (p.xiii) with prescribed identities, something Butler explicitly connects to 'democratic' life (Butler, 2011 :xiii).
During the fieldwork, the researcher kept an ethnographic journal, which contains approximately 70,000 words of text and a notebook of longhand impressions, written during quieter periods in the office and in the evenings.The journal was enriched by more immediate impressions and responses to unfolding events, captured through the notes, voice memo and camera functions on her mobile phone.In addition, 79 hours of meetings and events were recorded, and 28 interviews were conducted, each lasting approximately 1.5 hours.Participants comprised staff members of WRC ( 18), allies of the organization (6), government employees (2) and international organization practitioners (2).
Non-WRC interviewees were identified through their narratological connection with the actionpeople the first author encountered within unfolding events.The researcher supplemented these with media created by herself and WRC: videos, reports, posters, photographs and public announcements.

| Data analysis
The analysis focused on the multimodal, linguistic and visual, incorporating the enunciations of research participants, but also the reflexive and embodied account of the first author from her ethnographic journal.We explored the agency at play when the linguistic engaged with the visual, enabling us to interpret the embodied performative work unfolding in the scene.All data was generated in the Montenegrin language and analysed in its original form, especially important for textual discourse analysis, as the translation to English would alter not only the syntagmatic configuration of text but would also distribute the performativity of the language differently (Dreyfus, 2017).
The first stage of analysis was performed throughout the fieldwork, an iterative process of action, observation, reflexive writing, interpretation and re-writing.The engagement of research participants was crucial, involving substantial dialogue about emergent ideas, prompts and themes, which enabled a continuous co-construction of knowledge.Whenever the first author thought she was circling a significant theme in her preliminary analysis she would make this train of thought an item for discussion at a regular Monday morning staff meeting and within the next round of research interviews.This strategy enabled more egalitarian and co-constructive knowledge generation, where decisions regarding analytical categories were negotiated, and where participants had direct insight into the data analysis process.Such an approach was informed by the democratic tenor of the research topic, so that its participative content was inseparable from its form.
Within this analytic process, while the authorial voice of the second and third authors appears to be less powerful due to their lack of direct immersion in the field, their (relative) disembodiment also provided different benefits.
This was partly due to their 'outsider' status; they were not from Montenegro, and neither were they intensely immersed in the scene.Unlike the first author, who was both researcher and subject and therefore doubly burdened (and blessed) by 'the dilemma of closeness/distance' (Alvesson, 2009, p. 162), the second and third authors were burdened by distance alone, so their propensity to challenge, to take an alternative, less ethnocentric, and more plural perspective on the study was (relatively) easy.In any form of co-production it is difficult to disentangle conversations, revisions, comments and reshaping of data categories from what emerges as a final and conceptualized 'output', for the sum of the parts always exceed themselves.Suffice to say, that each voice was/is different, but not entirely distinct.
These analytical encounters led to the generation of three themes of exposing, citing and inhabiting, which were fed back and refined through ongoing cycles of reflective dialogue with participants.Using the data generated within these themes, we embarked on a second stage of analysis, adopting principles of multimodal and Butlerian discourse analysis (Harding, Lee, & Ford, 2014;Riach, Rumens, & Tyler, 2016).The performativity of the visual and linguistic material was analysed to explore the interface between the body and language, looking at how it supported or inhibited solidarity practices in relation to dominant discourses.
Within this stage we followed three steps.The first analysed the argumentation approaches of participants, which enabled us to understand the meaning and purpose attributed to solidarity building.Of interest therefore were argumentative enunciations made by participants within speech but also through visual artefacts such as placards, flags and posters.Adopting a Butlerian approach to analysing argumentation, we sought to analyse how meaning was grounded through the repetition of arguments and the ways in which some repetitions (in language and visually) went awry, creating alternative possibilities for constructing solidarity.
The second step analysed individual grammatical units of speech and how they related to one another and the arguments constructed.We focused on the use of verbs, nouns, pronouns, adjectives, prepositions, conjunctions and interjections (Harding et al., 2014;Lirola & Chovanec, 2012).As a third step we conducted a visual analysis and explored the ways in which the visual overlapped and co-constituted the linguistic.Visual resources are particularly important for exploring 'identities, values and activities' (Machin & Mayr, 2012, p. 15), as they offer a series of associations that can be further interrogated (Linstead, 2018).
Our performative visual analysis focused on the objects, settings and people presented in visual material (e.g., images, posters, banners) and the relationships they were situated within (Machin, 2016).As someone who was immersed as a participant, the first author collected but also produced artefactsshe took photographs but also contributed to the making of placards and broader discussions about the aesthetic strategies of the protests.In interpreting visual artefacts, we particularly focused on the performative work of perspective, setting, colour, tone and focus (Machin & Mayr, 2012) and what these can 'afford' in a particular context (Alcadipani & Islam, 2017).Such an analysis was not only a matter for a more detached reading after the fact but also one of exploring through ethnographic writing soon after the event, the disruptive and embodied experiences of the first author at the heart of unfolding eventson the streets, in the heat of protest and activity.

| FINDINGS: EMBODIED AGONISTIC SOLIDARITY BUILDING
Three dimensions of embodied agonistic solidarity building are analysed: exposing, citing and inhabiting, each denoting a different aspect of the practice that unfolds through contest.These draw on different kinds of data for their evocative powerethnographic writing, visual material, interviews and informal conversationsin an attempt to embed the reader within the rich tapestry of events as much as possible and to provoke an embodied reading of the text.

| Exposing
This dimension highlights the exposing of one's body to the precarious and embodied experiences of others, facilitating a relational emergence of agency (Butler, 2006(Butler, , 2015)).Rather than seeking the safety of a more bounded and compartmentalized engagement, practitioners become increasingly aware of the particular nature of their own stances through an embodied discovery of others for which there is 'no ready vocabulary' (Butler, 2006, p. 22).
To illustrate, we draw on an excerpt of exposing from the journal entry written after the primary researcher's first encounter with the protest of mothers.Collectively, WRC had decided to walk amongst the mothers to make connections and gather information: I stepped into the crowd confident of my view that the amendments to the law on social and child protection were outright wrong.In my mind I was rehearsing a script for conveying this in a friendly manner in case I was asked.However, squeezing between grids of raised arms gripping banners with calloused hands, which perfectly framed their determined but weary faces, my confidence started to erode.Suddenly, all the clues indicating the kind of lives many of the women here lived, rushed to my attention like a meteor storm: wrinkles on relatively young faces matching those on the more mature, in all but the depth of creases, worn out shoes and jackets, no jackets at all, swollen bags underneath eyes, an intricate web of tiny cuts and scars on their knotty hands induced by years of manual labour, solemn and exhausted facial expressions.
The rationalised political stance I came to the protest with started to wobble and became overpowered by feelings of compassion towards these women.
My heart raced with panic when I noticed a woman smiling and moving in my direction as I realised that the script I was rehearsing was rapidly dissolving in the face of these new realisations.The women I had visualised in my mind previously did not wear the marks of living in hardship.I was aware that I would never be able to explain to any of these women why I was against the benefits, as saying that would mean that I was standing against them putting food on the table for their families, providing books for their children or being able to pay their utility bills.I made small-talk with the woman, during which I learned that she was a mother of four and wife to a partner who suffered permanent spinal damage performing perilous construction work, and how benefits had been the only source of income for them in the past months.The realisation of the burden of her hardship prompted me to re-evaluate and reframe my political stance.
When I answered in the negative to her question about whether I had children, an expression of puzzlement briefly washed over her face before she offered a form of empathetic understanding, saying that: 'It's cruel that young women can't afford motherhood.' Afraid to say anything that might give me away as an outsider, an 'enemy', I simply thanked her and continued to move amongst the crowd, further feelings of guilt creeping in … I became overwhelmed by feelings of anger, injustice, betrayal … My stance was wellinformed, thought-through, feminist(!), but it felt wrong.
The excerpt offers a glimpse into an emotional rollercoaster, triggered by engaging with difference (Mouffe, 2013, p. 5).In a short span of time, the researcher transits from being confident about her political stance to the point where she feels it is wrong.There is little argumentation offered to support this transition; the process is predominantly fuelled by the embodied experience of others' difference.
There is an abundance of visual signifiers within this extract, for example: 'calloused hands' and 'wrinkles on young faces' confront the researcher as she joins the protesting mothers.These signifiers propel the initially confident investigator on a tumultuous emotional ride: from being confused to experiencing 'compassion' and then 'panic' when her political stance begins to 'wobble'.The panic then gives way to feelings of 'guilt', 'anger' and 'betrayal' before she experiences her political stance as 'wrong', without rational arguments to support this.Such embodied experience suggests that mere exposure to visual signifiers prompts re-evaluation, which further implies that dialogue between the protestors and the researcher had begun -'assembly is already speaking before it utters any words' (Butler, 2015, p. 156).
Furthermore, such embodied 'dialogue' has an antagonistic element, as the mothers' visual signifiers disrupt the researcher like a 'meteor storm'.The 'rehearsing' of arguments in the researcher's mind does not withstand such an 'attack' and her political stance becomes a 'dissolv[ed] … script'.The more it dissolves, the closer these unlikely allies seem to becomea connection through 'compassion' and 'empathy' is being made.
The short verbal exchange between a mother and the researcher is also rich in visual signifiers, expressed via an evocative combination of adjectives and nouns.'Spinal damage' and 'perilous construction work' hint at the hardship of supporting six family members on meagre benefits.These signifiers help to dissolve the researcher's political stance, whose priority switches to concealing her identity as someone who does not belong, as expressed through potent nouns such as 'outsider' and 'enemy'.Such a strategy implies opening, rather than closing, the possibility for a potential alliance (Mouffe, 2009a, p. 13).It also suggests that even though neither political stances nor terms of alliance have been articulated, as in conventional debate, some basis is formed through an embodied responsiveness; a simple, albeit uneasy, recognition of mutual vulnerability.
The dimension of exposing provides insight into how a basis for solidarity building is created between unlikely actors, through embodied 'dialogue' prompted by visual signifiers.Such dialogue, albeit driven by compassion, involves an antagonistic component where practitioners' stances are shaken under the stream of visual 'demands' articulated through bodies.

| Citing
This dimension is strongly informed by Butler's (2004, p. 52; see also Butler, 2015Butler, , 2016) ) account of the reproduction of norms, where 'they are invoked and cited by bodily practices that also have the capacity to alter norms in the course of their citation'.Here we build on the notion of 'citing' to explore the ways in which agency is asserted in visual and bodily ways, which we posit as implying collectivityan absorption, visual display and reappropriation of others' subjectivity into one's own.This dimension is distinct in the sense that participants who encounter and develop solidarity together enact and display it visually in public and through such exhibition further cement their sense of solidarity.These acts are an agentic and semiotic assemblage of available signs and symbols, which craft a collective purpose and solidarity amongst diverse subjects that crosses over from the linguistic to embodied.
Practitioners wanted to support the mothers, despite their differences, but no one was certain how this could be enacted in practice.For a number of weeks, the protestors and WRC remained locked in a form of political stasis, where dialogue between representatives of the two groups jarred.During this time the mothers simply wanted their benefits restored, focusing on their immediate financial hardship and socially privileged position as mothers, while WRC wanted to protect the rule of law by drawing attention to the point that the government could not guarantee rights and then take them away.Moreover, WRC believed the mothers were economically impoverished not only because their benefits were revoked, but mainly by the suspicious privatization initiatives and corrupt investments from government that left them without any means of supporting themselves.
Divergences between WRC and the mothers were not constituted through words alone.Some of the tensions can be gleaned from Figure 1.In the photograph, taken in the early stages of the protest, a woman holds a sign saying: 'Mothers are the Law', a pun which roughly translates as 'mothers rule' and denotes the citing of a powerful social norm (Butler, 2015, p. 63)the importance attributed to the status and authority of 'mother'.The norm of motherhood as privileged becomes even more pronounced through the underlined red lettering, as if the words could have been sanctified in blood.The photograph's staging re-enforces the message of the 'sanctity of motherhood' discourseindeed this was a consistent theme in the first weeks of the protest.The gaze of the mother is directed at the vieweran engaging look (Machin, 2016).When connected with the 'sanctity of motherhood' discourse, which is so prevalent in Montenegro, we can plausibly assume that the anticipated response is an act of validation (recognition) on the part of the viewer.speeches, which brought them and the practitioners closer together, although their differences were not elided.
Figure 2 uses the same staging techniques as seen in the previous photograph but delivers a subtly different message.
In this photograph the sign says: 'We only ask for what belongs to us according to law'.Here the message of the mother to the viewer is a plea for empathy: the feeling that one has been deprived of one's rights in law by an antagonist government, and contains an expression of commonality, 'like you'.The photograph demonstrates a shift from the particular subjectivity 'maligned mother' to a more inclusive one of 'maligned mother who has been wronged just like you'; a crossing that 'both binds and differentiates' (Butler, 2015, p. 77).Her positioning is conveyed neither through word nor image alone, but through an accumulation: the mother's body is brought to bear on the visual of the sign and the words communicated.The mother cites the words of WRC, but also through her bodily presence (she looks run down), she enhances and reconfigures the words into a new political stance.
The extent of the shift through citing can be seen in Figure 3.Here we see a mother in one of the later protests holding a flag designed by a progressive and feminist NGO (Anima), denoting love and solidarity.The citing that occurs here is far more general and broadly progressive than the original language of the mothers and carries a more encompassing and perhaps impactful message than had a practitioner been holding it.Furthermore, the photograph itself denotes connection between diverse people and groupsthere are many people in frame and the angle of the photograph is downwards to emphasize the number of people present: their individual features are less important than the mass of people and the prominence of the flag they congregate beneath.
Having lost their 'infrastructure of support' (Butler, 2016, p. 19), the mothers discover an alternative infrastructurein the form of building an alliance with practitioners.This solidarity is not just underwritten by circumstantial or technical means but encompasses the symbolic and the embodied.It is displayed and enacted through the appropriation of new signifiers from practitionersinfrastructural and relational symbols of solidarity, democracy and feminismin order to augment their subjectivity.Such augmenting is enacted through a mix of bodies, texts and visual signs and it is difficult to conceptually disentangle any of these.
can be viewed at wileyonlinelibrary.com]

| Inhabiting
This dimension denotes the embodied acts of inhabiting the experiences of othersfeeling their hardships through the body and therefore performing solidarity through the body (Butler, 2006(Butler, , 2016)).This is a relational aspect of solidarity building, whereby the subjectivities of those represented and the representing practitioner are influenced and changed in the process.It is also a process marked by care for the other's vulnerability through embracing vulnerability oneself.
The episode used here draws on data generated during the hunger strike of mothers, which lasted for 11 days.
Mina, who joined the mothers in their hunger strike, is a seasoned practitioner with a high profile and long history of participation in various protests and street actions.When the researcher asked her if she could explain what drove her decision to join the hunger strike, she responded: We were against these benefits … but these women were used and deceived … it just became clear to me that they are all alone … sustaining blows from everywhere … I wanted our civil society to stand in solidarity with these mothers, and not only us.I hoped that others would recognise another political dimension of the protest, that the mothers were not alone, that the civil sector supported them.This effort to fortify the protest, to add an additional buffer to mothers … I wanted to be there as an activist from the civil sector, but I also hoped this engagement would motivate other civil society members to support the mothers.I wanted them to understand that what these women were doing was relevant to the entire society, and I wanted us to sustain their engagement.I put my body out there as a political tool, if you like … First, I wanted to prove to mothers that we meant it, that our support was sincere, that we were in it together.Second, I wanted to soften the blows they were suffering from everywhere, you know, to erect another wall of support so to speak … So I placed my body on the crossroads of political streams: civil sector activists who were against the hunger strike … because they knew that the corrupt politicians wouldn't give a damn about it, but this was the mothers' decision, so if we were to support their cause we had to respect their means of protesting.Then there were institutions with their rotten politicians and media who tried to diminish the efforts of women where we wanted to amplify them and, finally, I wanted to prove to these women that I was on their side, our side, to break that wall of suspicion.These women represented everything worth fighting for … they were a manifestation of our society's hardships and they had to endure these because of our corrupt government, suffering the biggest brunt of it all their lives … I didn't want them to suffer alone and fight alone.Their problem was my problem, our problem.
Mina suggests that her decision to support the mothers is grounded in a sense of solidarity for those who have been abandoned, rather than a political decision to either support or oppose the benefits.In the face of the present political reality, she judges that mothers have been 'used and deceived' and exposing and correcting the injustice becomes a priority, the basis upon which she frames her actions.She assumes 'adversarial' relations with the mothers, albeit rooted in solidarity, and, by default, an 'enemy' position towards the state (Mouffe, 2013).We learn that mothers are not only alone, but that they are also 'sustaining blows from everywhere' in a myriad of ways, which partly explains Mina's desire to protect them.Throughout the rest of the excerpt she details what she hopes to gain by participating in the hunger strike: to navigate between the three political 'streams' (mothers, the broader civil sector and institutions) in an attempt to secure trust, increase participation and 'fortify' the power of the protest.
The ways in which Mina uses pronouns is insightful for understanding her stance.She shifts between the single person 'I'/'me'/'my' and collective pronouns 'we'/'us'/'our' when she speaks about both the protestors and civil sector practitioners.This movement may denote a confused sense of belonging (Harding et al., 2014).Yet, in light of the argumentation offered, this crossing back and forth between different pronouns more likely indicates a gradual inhabiting of the subjectivity of others, moving towards a more complete solidarity.In the face of 'disappearing infrastructures' (Butler, 2016, p. 21), Mina uses her body to blur the lines between the mothers and practitioners in order to demonstrate the strength inherent in the absorbed subjectivities.Her bodily performativity enacts a new subjectivity where the 'I' and the 'we' are entangled, which in Mina's view provides strength to the protest.
This dimension provides insight into the process of solidarity building where practitioners inhabit another's subjectivity in order to strengthen an assemblyan alliance between mothers and practitioners.Such inhabiting is relational, accomplished by participating in the painful experiences of others, a process which enables collective agency to be enacted.Moreover, the process of inhabiting enables a new political subjectivity to emerge: the protesting political subject is no longer simply a grieving mother who wants her money back, but a new, strengthened political subject who can amplify and expand her demands to include breaches of the rule of law.This dimension is an important aspect of solidarity building, as it enables the assembly to persist and new demands to be made against the 'enemy'the corrupt government.

| DISCUSSION
This inquiry has presented feminist solidarity as a contested and embodied democratic practice hewn together between diverse and unlikely alliesprotesting mothers and civil sector practitionersand against the actions of a corrupt government and its agents.Three dimensions of an embodied form of agonistic feminist solidarity were explored.Exposing was posited as a preverbal form of agonistic solidarity building where the subject engages in a visceral and embodied re-evaluation of their political stance and identifications.Citing was elaborated upon as the formation of assembly through public displays of amalgamated subjectivities, drawn from differences.Inhabiting was articulated as a form of sharing pain and vulnerability in order to build an expanded, more empathetic and compassionate form of solidarity within a new political subject.In what follows, we return to the solidarity, agonism and embodied performativity literature to elaborate on our contributions and to suggest future paths for the study of feminist solidarity.
Our first point concerns diversity in solidarity practice and research.We share the optimism of Vacchani and Pullen ( 2019) with regards to the potential of feminist solidarity to enact major systemic change; a solidarity that is not based on a nostalgic or naive faith in the capacity of the category of 'woman' or 'global sisterhood' to address the inequalities and oppressions suffered by each social and identity group.This ethnographic study reaffirms other studies (Bonfiglioli, 2016;Cranford, 2012) in showing how stark differences reside within the category of womanin our case, a conservative view of woman as mother and a more progressive view of woman as oppressed at the intersection of a number of identities.
Yet, instead of collapsing, erasing or tolerating difference through insisting on commonalities, we view ineradicable difference and diversity (Mouffe, 2009a(Mouffe, , 2013) ) as essential for feminist solidarity projects, both in practice and in academic study.Just as the scene of solidarity practice can be enlivened by agonistic and embodied forms of conflict, and generative engagement of difference, so it may be the task of scholars to seek out diverse conceptual and methodological approaches capable of enriching understanding.Conceptually, we place the agonism of Mouffe and embodied performativity of Butler in a critical dialogue with one another, together with practice.In doing so we reject the dualistic proposition that working with difference in solidarity can be reduced to a matter of language or body, but rather it encompasses these in agonistic tension and as inseparable in practice.
Part of our contribution in theorizing how solidarity works with difference in practice is to introduce the frame of democracy.Conceptualizing solidarity as a form of democratic practice exceeds counting bodies or majority rule and becomes a participative endeavour where bodies engage with one another in fluidity, contest and responsiveness.In our study, making sense of practice as an agonistic form of democracy helps us see this process between mothers and NGO practitioners as confrontational, one where conflict helps forge alliances.Nevertheless, an agonistic account of democratic practice also entails accepting 'exclusion' (Mouffe, 2009b, p. 551) as an inevitable part of forging solidarity.Such exclusion is not free licence to denigrate the interests and views of participants but an acknowledgment of the need in any collective for an external force or actor against whom the group may constitute its emergent agency.In our study both mothers and practitioners might have excluded one another and continued their work independently, but instead, through struggle and discovery, constituted an alliance in negative relation to a government actor posited as malign and corrupta shared 'enemy' (Mouffe, 2013, p. 7).Exclusions are not prohibitive or final and a democratic focus helps us see how such groups develop, as positions and identifications shift over time.
A focus on diversity and difference in feminist solidarity should also carry over into methodological considerations.Here we sought to engage with the diverse identifications at play in the scene through adopting equally diverse research methods.Solidarity is a rich, complex and multifaceted enterprise, encompassing language and the senses, that develops, twists and turns through time.We affirm an embedded and ethnographic approach to studying feminist solidarity as a vital means of evoking the ongoing struggle and sensorium, as participants confront and challenge (expose), mimic or mime (cite) and enrich and amalgamate (inhabit).Providing a rich and immersive account of solidarity therefore means presenting its embodied, visual and textual dimensions, in tension and in generative dialogue.In our study participants made sense of their experiences in words, for example, when the researcher sought to comprehend encounters at a protest after the fact through a journal.Yet ultimately such sensemaking through language was inseparable from the visual and the embodiedin this instance, through the mingling of feelings and linguistic imagery.Elaborating on the sensorium of feminist solidarity remains an important task for scholars and we view great potential in expanding an aesthetic interpretation, encompassing the auditory, tactile and visual.
Our second point of discussion concerns embodiment in relation to feminist solidarity (Butler, 2015(Butler, , 2016;;Hemmings, 2012;Vacchani & Pullen, 2019).Our findings suggest that solidarity endeavours are not forged solely through debates, discursively mediated consensus or deliberation but also preverbally, where claims are made upon the flesh of participants, without words being uttered (Butler, 2015).In exposing, we saw how the researcher's political stance was unsettled through her embodied response to the bodies of women in need, a felt contest of visual signifiers that occurred prior to a more rationalized sense being made of the experience through writing.Similarly, in both inhabiting and citing, women respond to one another through their bodies -through standing together and suffering togetherbefore such solidarity is communicated in language.This embodied engagement, when viewed as enmeshed in agonistic practice, helps us see how 'dissonance' (Hemmings, 2012;Vacchani & Pullen, 2019), experienced first through the body, can foster collective agency, as women express their shared sense of anger, grief and joy in assembly.
Drawing on Butler's (2015) account can help us interpret how participants in the study rejected their socially prescribed identities through enacting new subjectivities collectively in assembly.They did so by putting their bodies to use in unexpected ways (Butler, 2011) the mothers were expected to be quiet, subaltern servants of the ruling patriarchy and the NGO practitioners to be involved in behind-the-scenes care and advice, or policy advocacyyet they each defied these norms through their embodied acts.However, we can develop this insight by demonstrating how assembly, as a form of enacted solidarity, is not only a matter for bodies but of the intermingling of bodies, visual signifiers and text.In practice it is impossible to separate thesebodies mould the visual of signs and the message of a text, and vice versa.Poor conservative mothers with deep creases and knotted hands insisting on the rule of law carries more weight than a similarly conveyed message by NGO practitioners and, likewise, NGO practitioners starving for the mothers' cause extends the visual and discursive chain of solidarity.
The embodied aspects of solidarity are interpreted as experiences of felt and shared vulnerability (Butler, 2006(Butler, , 2015(Butler, , 2016;;Tyler, 2019) and we concur with Tyler's (2019, p. 61) claim that 'social solidarity emerges from mutual recognition of our embodied, relational multiplicity and shared intercorporeal vulnerability'.
Through joining the mothers in hunger, for example, Mina immersed herself in the suffering of her new allies.
Such acts help in the cultivation of empathy through breaking down some of the distance between participants, allowing them to 'echo' (Lyshaug, 2006, p. 99) the pain of others (Vacchani & Pullen, 2019, p. 36).Butler's (2006Butler's ( , 2015Butler's ( , 2016) ) conceptualization of shared vulnerability and relational infrastructure, as moments of ethical connection, help us view the potential for allies to translate embodied moments of dissonance into shared expressions of dissatisfaction with their lives under neoliberal conditions (Hemmings, 2012;Vacchani & Pullen, 2019).
Yet there is a danger here that shared vulnerability, compassion and empathy are interpreted in benign and overly positive terms.In our study such shared vulnerability was ridden with antagonistic moments at the level of the subject: the first author experienced extreme dissonance as she encountered the vulnerability of others through the visual signifiers of their bodies, and Mina experienced physical pain as she starved her body.We can here bring together notions of agonism with those of a relational infrastructure to see how agonism can be felt and enacted in embodied waysbodies learn about other bodies' vulnerability and shift their identifications and shared commitments.However, there is an ever-present danger that such embodied agonistic engagement may tip over into violent antagonism, tearing at tissue and causing lasting damage to participants.Issues such as activist and practitioner burn-out and the ethics of embodied solidarity actions become relevant and remain fruitful avenues for future research.

| CONCLUSION
The purpose of this study has been to theorize a form of feminist solidarity practice through a democratic and embodied perspective, to reframe difference and contest within solidarity as a resource, rather than problem.
This has been demonstrated by placing agonistic and embodied performative theory in conversation with ethnographic data from a study of solidarity in practice, within a patriarchal and corrupt society.Working with and through difference was posited as a democratic endeavour because this frame surfaces contestation as generative of solidarity building; where bodies, visual signs and language are in conflictual conversation but also productively enacting collective agency.Democratic theory was allied with embodied performativity to gain insight into preverbal, embodied aspects of this work.We theorized embodied agonistic solidarity as unfolding through SMOLOVI C JONES ET AL.
three dimensions.First, the exposing of one's body to the hardship of others to enable alliances between unlikely allies to form.Second, citing as a means of drawing in the symbolic resources of others and publicly affirming them.Third, creating a new political subject through inhabiting, which strengthens solidarity and enables assembly to grow and persist.We acknowledge that these dimensions far from exhaust the possibilities for solidarity and in the democratic spirit of the study welcome new interpretations and contestations that can further extend our knowledge of how solidarity unfolds in practice.
Gradually a resolution was reached about the legitimacy of the protest between various subjectivities, where the compassion of practitioners aligned with their political commitments to more radical forms of equality.The mothers' subjectivity moved towards the larger anti-authoritarian position of WRC.As a result of numerous conversations between the two groups, the mothers began to incorporate some WRC rule-of-law argumentation into their F I G U R E 1 Prelevi c (2017a) [Colour figure can be viewed at wileyonlinelibrary.com]

F
I G U R E 3 Prelevi c (2017b) [Colour figure can be viewed at wileyonlinelibrary.com]